The long-term objective is to determine the feasibility of applying innovative molecular approaches to improve cancer diagnosis in effusion specimens. Ascites (peritoneal) or pleural effusion is a common clinical presentation in patients with neoplastic and a variety of benign diseases. Diagnosis of these effusion specimens has paramount importance for clinical oncologists to best manage patients. Traditionally, cytological evaluation based on morphological features of cells is performed to diagnose malignant versus benign effusions and to determine the primary sites of carcinomas in malignant effusions. Unfortunately, such morphological examination is subjective in nature and often fails to detect malignant cells or identify the origin of the carcinoma in a malignant effusion. We hypothesize that 1) detection of tumor-associated genetic alterations (allelic imbalance and increased DNA integrity) in tumor-released DNA and the secreted protein markers (HLA-G, kallikrein 5 and kallikrein 8) alone or in combination provide novel molecular approaches to increase the sensitivity and specificity in diagnosing malignant ascites and 2) the primary site of a malignant effusion can be precisely predicted using a gene expression based cancer classification map. To test both hypotheses, we propose to detect multiple molecular genetic changes in tumor-released DNA and measure the levels of secretory HLA-G, kallikrein 5 and kallikrein 8 in the ascites supernatant. We will establish a cancer classification map based on the expression of a limited set of genes to predict the cancer types in malignant effusions. Translational researches aiming at the molecular diagnostics for cancer are urgently needed in clinical oncology because of the limitations in traditional morphological approaches to diagnose cancer in effusion specimens. The overall significance of this proposal is to develop innovative molecular assays to assist cancer diagnosis in patients who present effusion, a common clinical presentation in cancer patients. Although these patients often present at a late stage, outcome can be positively affected by accurate diagnoses followed by appropriate therapeutic regimens specific to different types of carcinoma. The results from this proposal will likely introduce new paradigms to molecular cytology and provide oncologists a new perspective in managing their patients.